Label Probabilistic Routing Recommendation Algorithm Based On Node Interest In DTN

Delay-Tolerant Network(DTN)refers to a type of network that operates under extreme conditions for communication.In many real-world applications,there are energy shortages or a series of uncontrollable interference factors that make traditional network protocols ineffective.As a new type of network architecture,DTN can better adapt to these challenging environments and meet the communication needs of these applications.Currently,more and more application scenarios are deploying and using DTN,such as mobile vehicular networks,interstellar networks,underwater sensor networks,and space optical communication networks.The most important aspect of these applications is the design of network protocols.Typically,a physical network relies on multiple protocols,where lower layer protocols provide basic functionality and higher layer protocols enrich these functionalities.Among these basic functionalities,routing,as a network layer protocol,has always been regarded as a hot research topic,as it determines the method and path of data packet transmission.Therefore,how to design efficient routing functions for DTN has become one of the key tasks.In DTN routing algorithms,Prophet is a highly representative algorithm that solves the blindness of data packet delivery.It measures the probability of nodes delivering packets by passing probability values,and selects relay nodes with higher probabilities of reaching the destination node based on this information to transmit messages.However,this algorithm only forwards the data packets to a small number of nodes,leading to an increase in message transmission delay.Additionally,the algorithm only measures the probability of reaching the destination node based on the encounter probability.In recent years,researchers have proposed many improved algorithms based on Prophet,but they still suffer from insufficient consideration of optimizing probability formulas,considering node social attributes,and facing sudden changes in node probabilities.Performance improvements are not significant,so there is still significant room for improvement in the Prophet algorithm.The specific research content and work are as follows:(1)The text is discussing a new routing algorithm called AESDP(adaptive exponential smoothing forecasting based on delivery probability of node)for delaytolerant networks(DTN)based on adaptive exponential smoothing prediction.Firstly,it analyzes the social properties of nodes and evaluates the predicted values of the transfer between nodes.It uses the predicted delivery probability value to help nodes selectively distribute message copies to reduce network congestion.Secondly,because the prediction model cannot adapt well to the dynamically changing network topology,an adaptive smoothing factor regulator is designed to adaptively adjust network model parameters,optimize relay node selection,and improve the overall network transmission performance.Finally,based on the ONE simulation platform and through cross-comparison experiments,it is confirmed that AESDP has a higher message delivery success rate than Prophet,NBAPR,and CARA,with an average of approximately 86.1%,59.1%,and 24.5%,respectively.(2)A label probabilistic routing recommendation algorithm based on node interest,LPRA-BNI,is proposed.LPRA-BNI analyzes routing recommendation strategies based on node interests and load balancing activation strategies for high-energy nodes from two aspects,namely node energy and accurate recommendation,to ensure the maximum utilization of network resources.First,the labels are used to describe the node location features,and then routing decisions are made based on node interest similarity.For nodes with similar probabilities,we use auxiliary information for further judgment and calculate the node interest factors and energy factors to form a recommended set of nodes.Highenergy nodes are assigned more tasks to avoid network black holes.The LPRA-BNI algorithm selectively transmits packets based on delivery prediction values,which not only improves the performance of the algorithm in terms of delivery rate and average delay but also effectively controls network overhead.Finally,based on the ONE simulation platform and through cross-comparison experiments,it is confirmed that the LPRA-BNI algorithm has the highest delivery rate.Compared to NBAPR,it increases on average by 61.07%,compared to CARA,it increases on average by 26.32%,and compared to AESDP,it increases on average by 1.3%.